11Health economics

Publication Details

11.1. Introduction

Health economic analysis allows decision makers to consider resource use alongside the benefits of a treatment in order to decide if it is good value compared with the next best alternative. Cost effectiveness analysis (with the units of effectiveness expressed in quality adjusted life years [QALYs]) is widely recognised as a useful approach for measuring and comparing different health interventions. Using the QALY as the final outcome allows one to measure the impact of health care in terms of how it extends life as well as how it affects health-related quality of life. Using this generic outcome allows different treatments to be compared using the same threshold for decision making.

For this guideline good quality published clinical evidence has been limited and therefore the benefits of treatment have been based on guideline development group (GDG) consensus. Where possible, economic analysis has been developed by working backwards from the NICE cost effectiveness threshold to determine what level of effectiveness would be necessary in order to find an intervention cost effective. This type of analysis does not give cost effectiveness results, but provides a framework within which to decide whether a treatment is likely to be good value in terms of NHS resources.

The NICE threshold is £20,000–£30,000 per QALY. For the treatment of spasticity it is the quality adjustment which is most important. Health-related quality of life can be measured in terms of the effect on domains of functioning and psychological wellbeing which focus on how well a person is able to carry out a full and meaningful life, such as mobility, self-care, ability to perform usual activities, pain/discomfort and anxiety/depression. The purpose of the treatments for spasticity considered in this guideline is to reduce pain, improve function and mobility, provide cosmetic improvements and prevent deterioration which may have resulted in loss of function. The view of the GDG was that improvements in these dimensions of health differ for each child or young person, depending on the extent of their impairment, their age and social context. Using this approach, whether a treatment ‘works’ for a particular child or young person can be determined by considerations other than the effectiveness of that treatment. For example, an orthosis may not have improved a child or young person's health-related quality of life, but it may have given them sufficient limb support to sit unaided. Where the goal of an intervention is not a measurable improvement in health-related quality of life alone, evaluating the cost effectiveness of specific treatments is a particular challenge.

For almost all the interventions considered in the guideline, published evidence of cost effectiveness was completely lacking. In the following areas, further analysis was undertaken to support the GDG's decision making:

  • physical therapy versus no active treatment
  • ankle–foot orthoses (AFOs) versus no active treatment
  • botulinum toxin (BoNT) versus oral drugs in combination with other interventions
  • continuous pump-administered intrathecal baclofen (CITB), including intrathecal baclofen (ITB) testing before CITB versus no ITB testing, and CITB versus oral drugs in combination with other interventions
  • orthopaedic surgery versus no active treatment
  • selective doral rhizotomy (SDR) versus no active treatment.

None of the analyses presented in this chapter follow NICE's reference case for health economic analysis because of the lack of evidence for effectiveness and because the GDG was not able to quantify the benefits of treatment in a way that could be used in an economic analysis using consensus values for unknown parameters.

In all of the topics considered for economic evaluation resource use and costs were quantified. Details of the methods used in relation to each review question are presented in this chapter. For each question the following are reported: review of published economic literature; description of resource use and costs; and conclusions of the analysis. A discussion of results was also included for questions for which a full health economic analysis was undertaken.

For many of the treatments considered in this guideline the GDG felt that the benefits to health-related quality of life outweighed the potential harms. Patient selection is important, particularly for the ITB pump, orthopaedic surgery and SDR, as only certain groups of patients are likely to benefit and treatment will not be appropriate for other groups. Patient choice is also important as their active participation, such as in physical therapy programmes and use of orthoses, is key to the success of several treatments.

Given the lack of published evidence, further comparative research is necessary to capture benefits in terms of function, pain, adverse events and quality of life. This further research should ideally be conducted using the EuroQol Group's EQ-5D instrument (a child-friendly version is available) or the Health Utilities Index (which was developed for children). Long-term outcomes are needed for the ITB pump, SDR and surgery as these are expensive and invasive treatments with associated risks. Also, the studies should be designed to allow subgroup analysis by severity of spasticity in terms of the Gross Motor Function Classification System (GMFCS) and limb involvement (hemiplegia, diplegia and quadriplegia). Studies should be designed to allow data on resource use to be collected to allow cost analysis. Cost effectiveness analysis comparing treatments for each subgroup would provide better information for decision making.

For each review question considered in the guideline, this document includes a health economic summary based on evidence and GDG opinion.

11.2. Physical therapy (physiotherapy and/or occupational therapy)

Health economic question

What is the cost effectiveness of physical therapy (physiotherapy and/or occupational therapy) in children with spasticity with or without other motor disorders (dystonia, muscle weakness and choreoathetosis) caused by a non-progressive brain disorder compared with no physical therapy?

Literature review

No published health economic evaluations were identified in the literature search conducted for this review question.

Further analysis

Introduction

The clinical evidence presented in this guideline was limited and could not be used to develop an economic evaluation for this review question. A ‘what-if’ analysis was considered. After much discussion the GDG came to the view that it would not be possible to quantify the mean benefits of physical therapy for the following reasons:

  • The guideline covers children and young people with considerable variation in impairment, from those with spasticity affecting how a single joint works to those with severe spasticity affecting all their limbs.
  • Therapeutic goals are personalised for each child or young person within the family and will change over time and in different contexts.
  • No research study was identified that quantified the mean benefit of physical therapy in a way that would be clinically meaningful.
  • The GDG was not able to come to a consensus view on what a single measurable health outcome would be for this group.

Given these factors, it was not possible to see how an evaluation could be undertaken that would provide meaningful results in which the GDG would have confidence. Therefore no economic evaluation could be undertaken. However, the resource implications of providing therapy were discussed by the GDG.

Methods

A cost description of the service was undertaken for the GDG using staff costs from Curtis 2011. This shows the costs of therapists providing care in different settings and for hourly sessions once, twice or three times a week. This was presented to the GDG to aid consideration of the costs related to providing physical therapy.

Resource use and costs

Table 11.1 shows the comparative costs for physiotherapy and occupational therapy across different NHS settings where only staff costs are taken into account. Client cost per hour includes the costs of overheads across the different settings but does not take into account the travel time required by community physiotherapists. The costs are very similar in all settings for both physiotherapy and occupational therapy.

Table 11.1. Cost description for physical therapy (Curtis 2011).

Table 11.1

Cost description for physical therapy (Curtis 2011).

Conclusion

Cost data for physical therapy have limited use without associated benefits. The cost of increasing physical therapy for children with spasticity could be significant but without knowing the benefits of increasing physical therapy we cannot know if it will be cost effective. GDG discussions identified that the therapist plays a key role not only in providing treatment, but also in assessing the patient and providing information to the parents or carers about treatment and ways to improve a child's daily tasks and activities.

11.3. Orthoses

Health economic question

What is the cost effectiveness of AFOs compared with no orthosis in children with spasticity and with or without other motor disorders caused by a non-progressive brain disorder?

Literature review

No published health economic evaluations were identified in the literature search conducted for this review question.

Further analysis

Introduction

As with physical therapy, there was limited good quality evidence for the effectiveness of using an orthosis compared with not using an orthosis. The outcomes of importance vary depending on the specific goal. The outcomes were reported for intermediate effects which, over time, could result in improvement in function and health-related quality of life. For example, the GDG considered that improved gait efficiency would contribute to subtle improvements in energy expenditure which could lead to increased activity and therefore allow more participation. As with physical therapy no research study has quantified the mean benefit of orthoses in a way that would be clinically meaningful. Also, orthoses are used in conjunction with other interventions, such as physical therapy and BoNT. It would be difficult to estimate the benefits of orthoses over and above these interventions. Again, as with physical therapy, it was not possible to undertake an evaluation that would provide meaningful results in which the GDG would have confidence.

Initially, a service description was developed with the assistance of Exeter University and the Royal Berkshire NHS Foundation Trust to potentially inform an economic evaluation.

Methods

Two orthotists were contacted who gave descriptions of the process for having an orthosis supplied and fitted. Costs were then applied to staff time for appointments and added to the cost of the orthosis. Data were presented to the GDG to allow the consideration of resource use and costs when making recommendations for orthoses.

Resource use and costs

Three appointments are required:

  • An initial assessment takes 20–30 minutes with a physiotherapist or occupational therapist and includes taking measurements.
  • A follow-up appointment to fit the orthosis takes 20–30 minutes and takes place about 2 weeks after the assessment.
  • A second follow-up appointment takes place to check everything is correct (this is usually offered only for a child or young person who has not had an orthosis before).

Orthotists start at band 5 and can work up to band 7 as a senior orthotist. Only a third of orthotists are employed by the NHS with the rest working for private companies. Using the cost per hour of client contact with a physiotherapist‡‡‡‡ (band 5 median) to represent the cost of an orthotist, the appointments will cost about £27 (40 minutes) to £62 (1.5 hours) to supply and fit an orthosis if the orthotist is employed in the NHS.

The cost of each AFO is about £120 to £300 (estimates from Exeter University and the Royal Berkshire NHS Foundation Trust). Lower limb orthoses are usually custom made, whereas upper limb orthoses can be products supplied from stock.

An orthosis needs to be replaced every 10–12 months or less, depending on the child's rate of growth. The straps on the orthosis usually wear out after about 12 months.

If an orthosis does not fit well and is uncomfortable then the child will not wear it.

Conclusion

As with physical therapy, the cost data for orthoses have limited use without associated benefits. Orthoses are used in conjunction with other treatments and it would be useful to understand the additional benefits of using an orthosis with each of the other interventions.

11.4. Botulinum toxin

Health economic question

What is the cost effectiveness of intramuscular botulinum toxin type A (BoNT-A) in combination with other interventions (physical therapy) compared with continuing on oral drugs with other interventions in children with spasticity and with or without other motor disorders (dystonia, muscle weakness and choreoathetosis) caused by a non-progressive brain disorder where there is no longer a beneficial effect from the oral drugs?

Literature review

No UK-based economic evaluations were identified from the literature search. A cost consequence analysis was identified for Australia (Houltram 2001), a cost minimisation analysis for Germany (Ruiz 2004) and a cost impact analysis from the USA (Balkrishnan 2002). Not enough detail was reported to adapt these analyses to the UK setting and, therefore, these studies are not discussed further here.

The clinical literature review identified that the evidence for the effectiveness of BoNT was equivocal. Cost effectiveness analysis was developed to consider what level of effect would be needed to find BoNT injections cost effective at the NICE threshold for cost effectiveness (£20,000 per QALY) compared with continuing on oral drugs.

Further analysis

Methods

The time horizon for the analysis was 1 year. The pharmacological activity of BoNT-A was assumed to last 3–4 months. There was limited evidence that showed no significant prolonged effect with repeated injections. The analysis considers one, two or three sets of injections over a year. Only costs relevant to the NHS are included in the analysis.

It is assumed that patients are referred fo BoNT treatment when oral drugs stop working, therefore the comparator will be continuing on oral drugs. All patients continue with physical therapy.

Resource use and costs

A cost description was developed based on service descriptions from Leeds Teaching Hospitals NHS Trust and Great Ormond Street Hospital. The BoNT service team comprises:

  • two consultants
  • a physiotherapist
  • an occupational therapist
  • a nurse
  • a registrar.

A new patient will have a detailed assessment performed by a consultant to determine their suitability for BoNT treatment.

The assessment includes:

  • clinical examination
  • video gait analysis
  • goniometry
  • measurement of gross motor function (not conducted for every child or young person)
  • agreement and documenting of treatment goals
  • completion of integrated care pathway paperwork
  • weighing the patient, obtaining consent and prescription of BoNT.

The administration of BoNT involves a day–case admission unless it is an inpatient referral.

All admissions require:

  • general examination to ensure fitness for sedation or general anaesthetic
  • parental consent for sedation or general anaesthetic and for injection.

The majority of injections are performed under sedation in the treatment room. Muscles to be injected are identified by a member of the BoNT team and marked, and a local anaesthetic is administered (AMETOP; tetracaine or amethocaine). A sedative is administered (oral midazolam at a dose of 0.5 mg/kg, maximum dose 15 mg). Patients who are old enough to cooperate, and are in agreement, will be offered entonox analgesia (nitrous oxide). This is usually combined with ethyl chloride spray anaesthesia. Entonox is administered by trained nurses.

A member of the BoNT service team performs the injections, using ultrasound guidance to locate the muscles. Once the child has woken and recovered they are discharged home. A handwritten discharge summary is completed and a dictated summary is produced afterwards by team members.

Follow-up appointments use the same assessments as pre-injection appointments. At Great Ormond Street Hospital there are two appointments at 3 and 17 weeks post injection; at Leeds Teaching Hospitals NHS Trust the follow-up appointment is at 6 weeks.

For the first appointment for the pre-assessment, the reference cost for a consultant-led face-to-face appointment was used, with the corresponding reference cost of follow-up appointments for any subsequent appointments (see Table 11.2).

Table 11.2. Costs of pre-assessment and follow-up assessment (NHS reference costs 2010/11).

Table 11.2

Costs of pre-assessment and follow-up assessment (NHS reference costs 2010/11).

The appointment for the injection of BoNT has a reference cost assigned – Torsion dystonia and other involuntary movements drugs band 1 (code XD09Z) – as it is a high cost drug (NHS 2006). The unit cost for 2010–11 was £321 (lower quartile £175, upper quartile £418). There is also a specialist uplift to tariffs for children of 60% (Department of Health 2011): if this is applied then the cost increases to £514. This reference cost will include all costs related to the procedure, the day case admission, drug costs and staff costs.

Standard care is understood to mean continuing oral drugs, assumed to be baclofen (Table 11.3). (Physiotherapy and occupational therapy costs are assumed to be the same for both treatment arms and so not included in the analysis.)

Table 11.3. Cost of standard care, oral baclofen, for 1 year.

Table 11.3

Cost of standard care, oral baclofen, for 1 year.

Costs to treat adverse events which are not transient are shown in Table 11.4. The GDG thought serious adverse events were very unlikely for patients receiving BoNT and so the baseline analysis was conducted without adverse events.

Table 11.4. Costs for adverse effects (NHS reference costs 2010/11).

Table 11.4

Costs for adverse effects (NHS reference costs 2010/11).

Outcomes

The clinical evidence from the trials was variable for reducing spasticity and optimising movement and function. The quality of life evidence only reported a significant benefit in the emotional role estimation. However, 66% to 81% of parents in one cross-over RCT rated BoNT treatment as good, very good or excellent.

The main adverse events reported in the literature review for this question were: incontinence; short term muscle weakness (Reddihough 2002); and one child with a history of epilepsy being admitted to hospital for seizure management shortly after injection (Russo 2007). In four studies grip weakness was reported (Boyd 2004, Fehlings 2000, Olesch 2010, Russo 2007). Other reports included nausea, vomiting, flu symptoms, coughing, soreness at injection site, respiratory infections, headache, fainting episodes, anxiety, depression, alopecia and fatigue.

Given the lack of clinical effectiveness evidence for BoNT, the benefits of treatment were estimated by the GDG. Using an EQ-5D UK time trade-off value set, descriptions of potential benefits were developed to help guide the GDG's decision making. In its discussion, the GDG identified the dimensions of the EQ-5D that would be most likely to be affected by BoNT treatment as mobility, self-care, ability to perform usual activities, pain/discomfort and anxiety/depression.

The conservative assumption for this analysis is that patients do not deteriorate if they continue on oral drugs.

Synthesis of costs and outcomes

Using the data from the clinical evidence and the costs above, a simple analysis shows the mean cost per person having two sets of BoNT injections in 1 year is approximately £2,000 (Table 11.5). This assumes injections are given in a day case setting and only one follow-up appointment is needed. The cost of the next best alternative (standard care for this analysis) is approximately £21 per year.

Table 11.5. Mean cost per child or young person for two sets of injections of botulinum toxin in 1 year.

Table 11.5

Mean cost per child or young person for two sets of injections of botulinum toxin in 1 year.

If two follow-up appointments are needed at 3 and 17 weeks after the first set of injections, then the average cost rises to around £2,600. If BoNT is given three times in a year then the costs increase to £2,800 per person per year.

Effectiveness threshold analysis

Given the lack of effectiveness evidence, the NICE threshold for cost effectiveness of £20,000 was used to calculate the minimum change in effectiveness that would be needed to find BoNT-A cost effective. The incremental cost of BoNT-A involving two sets of injections in 1 year compared with oral baclofen is:

£2,062 - £21 = £2,041

The incremental cost effectiveness ratio (ICER) is calculated as:

Incremental cost ÷ QALYs = ICER
£2,041 ÷ QALYs = £20,000 per QALY
£2,041 ÷ £20,000 = 0.1

So in order for two sets of BoNT-A injections to be considered cost effective compared with oral baclofen, a QALY gain of 0.1 would need to be achieved over the year.

The following descriptions were developed from the EuroQol Group's EQ-5D instrument to demonstrate what outcomes would be needed for the threshold for effectiveness to be reached. For patients who have moderate pain or discomfort, approximately 80% would have to experience no pain or discomfort if given two sets of BoNT injections during a year. For patients with extreme pain or discomfort, approximately 25% would have to experience only moderate pain or discomfort if given three sets of BoNT injections during a year.

For patients who have some problems with self care and some problems performing their usual activities, 80% would have to improve so that they had no problems with self-care or performing their usual activities if they have two sets of BoNT injections in a year.

The other way to consider the effectiveness of BoNT is as a prevention of deterioration: 80% of patients who have no pain are prevented from experiencing moderate pain if they are treated with BoNT after oral drugs fail; or 25% of patients who have moderate pain would be prevented from experiencing extreme pain; or 80% of patients who have no problems with self-care and performing usual activities are prevented from deteriorating that would result in them having some problems with self-care and performing usual activities.

Conclusion

There is uncertainty in this analysis as the clinical effectiveness evidence is variable. Only a small increase in quality of life is needed for this to be considered cost effective at the NICE threshold, and so even with uncertain clinical effectiveness it is likely that BoNT will be found cost effective to use. It seems from the clinical evidence that what is reported in the trials is not what the clinicians are looking for from BoNT in practice. Data on how BoNT treatment benefits children and young people in terms of mobility, self-care, usual activities, pain and discomfort, and anxiety and depression would be needed for further economic evaluation.

11.5. Intrathecal baclofen

Health economic questions

What is the cost effectiveness of an ITB test before receiving CITB compared with no test in children and young people with spasticity due to a non-progressive brain disorder?

What is the cost effectiveness of CITB in combination with other interventions (physiotherapy, occupational therapy) compared with oral drugs with other interventions?

Literature review

An economic evaluation set in the UK was identified in the literature search (Sampson 2002). The evaluation was clear and it was easy to identify the sources for the costs and effectiveness. It did not have a comparator intervention because the effectiveness evidence was based on case studies with no control groups. The model compared the costs of testing, implanting the pump and follow-up visits for 5 years (life of the battery for the pump) with the estimated benefits to quality of life.

The clinical effectiveness evidence used in the evaluation was taken from published studies identified in a literature search. Trials were included if they had more than one patient, an average follow-up of at least 6 months and included the following outcomes:

  • bedbound patients becoming able to sit in a wheelchair
  • patients who had severe difficulty sitting in a wheelchair being able to sit comfortably
  • wheelchair users improving their wheelchair mobility
  • wheelchair users improving their ability to transfer
  • wheelchair-bound patients becoming ambulatory
  • ambulatory patients improving their ability to walk
  • improved ability to perform activities of daily living
  • improved ease of nursing care
  • patients with skin integrity problems who showed improvements in these symptoms
  • reduction in spasm-related pain.

All the studies included patients with severe disabling spasticity no longer responding to oral medications and where the patients had already had a positive response to a bolus dose of ITB. Studies included children and adults with different causes of spasticity, but the results were reported for all patients together. The studies used a wide variety of outcomes. Functional and quality of life outcomes were generally not measured using standard scores.

The proportion of patients responding to treatment is an intermediate outcome. The authors translated intermediate outcomes into QALYs using EQ-5D scores based on the evidence identified in the review and supported by clinical opinion (see Table 11.6). The EuroQol Group's EQ-5D instrument measures health-related quality of life using five dimensions: mobility; self-care; usual activities; pain/discomfort; and anxiety/depression. Each dimension has three levels: no problems; some problems; and extreme problems. The patients were divided into three categories to estimate EQ-5D scores:

Table 11.6. Estimated improvement in quality of life based on proportion who experience changes in specified outcomes for three categories of severity of spasticity for adults and children (source:Sampson 2002).

Table 11.6

Estimated improvement in quality of life based on proportion who experience changes in specified outcomes for three categories of severity of spasticity for adults and children (source:Sampson 2002).

  • Category 1: bedbound patients experiencing severe spasm-related pain
  • Category 2: bedbound patients who were not in pain
  • Category 3: wheelchair users with moderate spasm-related pain.

The proportion of patients who would be expected to have no change in quality of life or changes in pain, depression, ability to sit in wheelchairs, ability to care for themselves and participation in activities of daily living are shown in column 2 of Table 11.6. These are based on combinations of the following specified outcomes from the literature search:

  • bedbound patients becoming able to sit in a wheelchair (66% of patients in the included studies responded)
  • improved ability to perform activities of daily living (73% of patients responded)
  • reduction in spasm-related pain (89% of patients responded).

Cost estimates published in the study were derived from three centres in the UK where the procedure was being performed. Benefits of the ITB pump were assumed to last 5 years as this is the lifetime of the pump's battery. Table 11.7 shows the costs reported from 1999.

Table 11.7. Change in mean cost of intrathecal baclofen testing and continuous pump-administered intrathecal baclofen over a 5-year period (Sampson 2002).

Table 11.7

Change in mean cost of intrathecal baclofen testing and continuous pump-administered intrathecal baclofen over a 5-year period (Sampson 2002).

The cost per QALY for each category of patients was:

  • Category 1 = £6900
  • Category 2 = £12,790
  • Category 3 = £8030

There was no comparator treatment in this study, therefore the results do not represent ICERs.

The conclusion of the study was that if the QALY gain was less than approximately 0.15 or if the cost of CITB was above £19,000 over the 5 year period then the cost per QALY would be greater than the NICE £20,000 threshold for willingness to pay for a QALY gain.

The published economic evaluation Sampson 2002 was used as the basis for developing a new model which looked at the cost effectiveness of both testing and implanting the ITB pump.

Further analysis

Methods

The costs of testing, implanting the pump and follow-up visits over 5 years have been taken from Sampson 2002 (see Table 11.7) and converted to 2010/11 costs (using the hospital and community health services pay and prices index uplift [Curtis 2011]).

As the model runs over 5 years, costs and benefits accrued after the first year are discounted by 3.5% for costs and 3.5% for benefits (1.5% tested for benefits). The perspective of this evaluation is from the NHS, therefore only includes costs and benefits relevant to the NHS.

A treatment pathway was developed with the help of the GDG in which additional elements of treatment were identified that were not included in the previous study (Sampson 2002). The main change to the published model structure was the inclusion of a comparator treatment. It was assumed that all patients would receive physiotherapy and so this was not included in the model.

In the model the following three comparisons were considered.

  • Children and young people considered suitable candidates have a pre-screening assessment and are tested before the pump is implanted. Patients who have a positive test will go on to have the pump implanted. Patients who have a negative test will have standard treatment.
  • Children and young people considered suitable candidates by their clinicians have a pre-screening assessment and get the pump implanted without a test dose.
  • ITB testing and pump is not available for any patients. Children and young people considered suitable candidates by their clinicians will continue to receive standard treatment.

The results of three small studies (reported in six publications) from the review of clinical evidence were combined to populate the baseline model parameters with a total of 117 patients. The studies (see Appendix L for further details) were:

Seven patients who had a positive test result but did not have the pump implanted have been excluded from the clinical evidence. Of these, six patients were excluded as they were ineligible to have a pump implanted, and there was one death unrelated to treatment.

Three of the children who had a positive test chose not to have the pump implanted. Although the exact reasons for these decisions are not clear from the studies, all three have been included in the economic analysis to reflect that some patients may choose not to have the pump implanted after the test (see Table 11.8).

Table 11.8. Clinical values and corresponding inputs for the test and no test arms of the model (percentage values rounded).

Table 11.8

Clinical values and corresponding inputs for the test and no test arms of the model (percentage values rounded).

In total, therefore, 110 of the 117 patients included in the studies from the clinical review were used to determine the risks for this analysis (see Table 11.8).

The GDG was asked to identify which adverse events reported in the clinical review for CITB should be included in the analysis. Adverse events related to baclofen are considered to be transient with low cost implications and minimal impact on quality of life.

The combined studies reported that seven pumps were removed and three patients required second operations to correct problems with the pump or catheter. One pump was removed due to lack of effect of ITB after a positive test. Table 11.9 reports the inputs for the model taken from the 110 patients included from the clinical review.

Table 11.9. Clinical values from 110 patients included from the clinical review and corresponding inputs for the CITB arms of the model (% values rounded).

Table 11.9

Clinical values from 110 patients included from the clinical review and corresponding inputs for the CITB arms of the model (% values rounded).

The following scenarios were included in the initial treatment pathway to be modelled, but no evidence was identified in the clinical review and so they have been removed from the final model:

  • technical failure of the pump requiring it to be removed
  • where no effect is seen or the effect is too small, then the dose is increased resulting in an additional follow-up visit
  • orthopaedic surgery (a possible outcome of using the ITB pump was thought to be delayed or prevented surgery).

Resource use and costs

Standard therapy was the continuation of physiotherapy and oral drugs (baclofen) for 5 years (see Table 11.10). It is assumed that patients with the ITB pump also continued with physiotherapy and so this cost was not included in either arm of the model. The updated 2010/11 costs for the ITB pump are reported in Table 11.7. Where the ITB treatment is unsuccessful and the patient has the pump removed they will go back to taking oral baclofen.

Table 11.10. Cost of standard therapy for 1 year.

Table 11.10

Cost of standard therapy for 1 year.

The review of the clinical literature for the guideline found evidence of adverse effects related to both the test and implanting the pump and so these have been included in the model. Both procedures require a hospital stay and involve injections or a catheter inserted into the spinal cord. There is a risk of infection which can be minor and easily treated, or a major infection such as meningitis. The costs of treating these infections are shown in Table 11.11. The costs are assumed to be the same whether the infection is due to the test procedure or the pump implant procedure. Meningitis is a major adverse event which can cause death or serious disability, with high associated costs. It is assumed that if a patient has a major infection the pump will not be implanted.

Table 11.11. Costs of treating infections due to test or implant procedures.

Table 11.11

Costs of treating infections due to test or implant procedures.

In the model if a patient develops a major infection during the surgery the pump will be removed. If the pump fails to work they will have the pump removed. For some patients the pump will be implanted but a problem is found that requires a second operation to fix. The costs of removing the pump or having a second operation to correct a problem are reported in Table 11.12.

Table 11.12. Cost of removing the pump due to major infection (2010/11 prices).

Table 11.12

Cost of removing the pump due to major infection (2010/11 prices).

Outcomes

No studies were identified that demonstrated the predictive value of the ITB test. Patients only had a pump implanted if the test was positive. After discussion with the GDG it appears that clinicians are able to predict which patients will benefit from ITB treatment from their clinical characteristics. The test is used to demonstrate the effectiveness of ITB to the patient and help decide treatment goals.

The baseline analysis assumes no improvement in quality of life for children and young people who have the pump implanted and this is the same effect as standard therapy, a conservative assumption to reflect that little good quality comparative evidence is available. It is assumed that staying on standard therapy resulted in no quality of life improvements, but also no deterioration.

The analysis was also run using the long-term quality of life effects from Sampson 2002 (see Tables 11.6 and 11.7).

Synthesis of costs and outcomes

The cost of care for a population of 100 children and young people was calculated for each arm (see Tables 11.13, 11.14 and 11.15). Table 11.13 reports the number of specific events (test results, adverse events) throughout the clinical pathway and the total cost for children who were tested prior to planned treatment. Table 11.14 shows the same data for children who were not tested prior to treatment and were identified as suitable to have an ITB pump based on clinical judgement alone.

Table 11.13. Population numbers, mean and total cost of intrathecal baclofen treatment with testing (N=100; figures are rounded from the model).

Table 11.13

Population numbers, mean and total cost of intrathecal baclofen treatment with testing (N=100; figures are rounded from the model).

Table 11.14. Population numbers, mean and total cost of intrathecal baclofen treatment without testing (N=100; figures are rounded from the model).

Table 11.14

Population numbers, mean and total cost of intrathecal baclofen treatment without testing (N=100; figures are rounded from the model).

Table 11.15. Population numbers, mean and total cost of standard treatment (N=100) (figures are rounded from the model).

Table 11.15

Population numbers, mean and total cost of standard treatment (N=100) (figures are rounded from the model).

In an economic evaluation a new treatment is always compared with another treatment or standard care. We are interested in the additional benefit of the new treatment above standard treatment and whether this incremental benefit is worth the additional cost. Using the baseline assumption of no improvement in health-related quality of life with CITB therapy, standard therapy should be chosen because implanting the pump incurs costs with no improvement in health. However, if the analysis is run using the quality of life outcomes from Sampson 2002 then using the ITB pump is cost effective compared with standard therapy in wheelchair users with moderate spasm-related pain (Tables 11.16, 11.17 and 11.18). The analysis suggests that implanting the pump without testing is less costly and more effective than testing first using these inputs, but the differences in the overall costs and benefits is small (£21,423 versus £21,370 per person, and 1.70 versus 1.71 QALYs over 5 years).

Table 11.16. Quality of life improvement scores for category 3 (wheelchair users with moderate spasm-related pain) from Sampson 2002 used in sensitivity analysis.

Table 11.16

Quality of life improvement scores for category 3 (wheelchair users with moderate spasm-related pain) from Sampson 2002 used in sensitivity analysis.

Table 11.17. Sensitivity analysis of incremental cost effectiveness results using quality of life outcomes from Sampson 2002 category 3 (wheelchair users with moderate spasm-related pain; benefits discounted by 3.5%).

Table 11.17

Sensitivity analysis of incremental cost effectiveness results using quality of life outcomes from Sampson 2002 category 3 (wheelchair users with moderate spasm-related pain; benefits discounted by 3.5%).

Table 11.18. Sensitivity analysis of incremental cost effectiveness results using quality of life outcomes from Sampson 2002 category 3 (benefits discounted by 1.5%).

Table 11.18

Sensitivity analysis of incremental cost effectiveness results using quality of life outcomes from Sampson 2002 category 3 (benefits discounted by 1.5%).

The incremental cost effectiveness results for all three categories of patients taken from Sampson 2002 are:

  • Category 1 (bedbound patients experiencing severe spasm-related pain) = £19,798
  • Category 2 (bedbound patients who were not in pain) = £10,691
  • Category 3 (wheelchair users with moderate spasm-related pain) = £12,431.

Discussion

There is considerable uncertainty in this model given the limited clinical evidence available to show the effectiveness of the pump. Only one RCT was identified with a very small study population of children and young people, but it was not a long-term study. Therefore the baseline assumption for this model was that the ITB pump would have no effect on quality of life and therefore could not be cost effective. This was tested in a sensitivity analysis using estimated quality of life scores from Sampson 2002. Using these quality of life scores the ITB pump may be a cost-effective treatment compared with standard treatment. The EQ-5D scores from Sampson 2002 were estimated by the authors, and included adults and children with different causes of spasticity and so these scores may not be representative of the improvement in children and young people who have spasticity as a result of a non-progressive brain disorder.

Given the lack of evidence for improvement in quality of life the model was used to calculate what level of effectiveness would be needed for the pump to be found cost effective at the NICE threshold of £20,000 per QALY. If the ITB pump improves quality of life by more than 0.26 each year for the 5 year lifetime of the pump, then implanting a pump would be cost effective by the NICE threshold. The effectiveness of testing is also uncertain. If there are no adverse events related to testing then the QALYs gained are equal for both the group having testing and the group not having testing. But there is still an additional cost related to testing patients of £1,837 per patient. Testing may avoid pumps being removed due to lack of effect, as patients who will not have a beneficial effect may be identified at the test stage. However, in this analysis the costs of testing outweighed the cost of additional surgeries required to remove the pump. There would be a quality of life decrement if the pump is removed but it would be short-term.

The GDG believes that testing is a valuable part of the treatment, as in some cases reducing spasticity can have a negative effect and then the pump would not be appropriate. Also the test would allow children and young people and their parents or carers to understand the effects of ITB and so make informed treatment choices and feel confident in giving consent.

The main costs are related to implanting the pump. Sensitivity analysis could be performed varying the costs included in the model. Given that standard therapy is so much cheaper than continuous ITB, the other costs included in the model, for example treating infections, are minor compared with the overall cost of testing and implanting the ITB pump.

The costs used in this model were uplifted from 1999 costs and these may not be representative of the true current costs. A document for the East Midlands Specialised Commissioning Group on the Commissioning Policy for Intrathecal Baclofen (accessed 12 June 2012) showed the costs of implanting an ITB pump for 1 year using 2009 costs (see Table 11.19).

Table 11.19. Cost of intrathecal baclofen pump for 1 year for paediatrics (uplifted to 2010/11 prices) (East Midlands Specialised Commissioning Group).

Table 11.19

Cost of intrathecal baclofen pump for 1 year for paediatrics (uplifted to 2010/11 prices) (East Midlands Specialised Commissioning Group).

The costs from Sampson 2002 were more detailed and so used in the model. Using the East Midlands costs, the overall cost with the test and including a 5 year follow up was £28,213. This is higher than the costs used in the model, but when tested in the model these higher costs did not change the direction of the results.

Conclusion

ITB is much more expensive than standard treatment and its clinical value is uncertain. This analysis illustrates the trade-off between the benefits of treatment, the risks and the costs. This is based on very limited, low quality data which suggests that the efficacy of this treatment, and the risks and adverse events associated with it, are not well known. A more detailed evaluation of the costs, benefits and risks of ITB require more long-term data, especially as this analysis suggests that ITB may be beneficial and cost effective in some groups of children, but not all children.

11.6. Orthopaedic surgery

Health economic question

What is the cost effectiveness of orthopaedic surgery in preventing or treating musculoskeletal deformity compared with no surgery in children with spasticity caused by a non-progressive brain disorder?

Literature review

No published health economic evaluations were identified in the literature search conducted for this review question.

Further analysis

Introduction

Given the lack of clinical evidence for the outcomes considered important for this question it was not possible to develop an economic evaluation. The guideline covers children with considerable variation in impairment from those with spasticity in a single muscle to children with severe spasticity affecting all limbs and appropriate surgical management procedures will vary between patients. Like the other interventions for spasticity considered for economic evaluation, the goals of therapy will be personalised for each child or young person and will change in different contexts. The outcomes of importance will therefore vary depending on the specific goal. As with physical therapy and orthoses, it was not possible to undertake an evaluation that would provide meaningful results in which the GDG would have confidence.

Methods

A cost analysis was requested by the GDG and is presented here. NHS reference costs from 2010/11 were found for various types of surgery. The cost of surgery varies depending on the limb and on whether it is minor or major surgery, and there is a 60% uplift to tariffs for children when surgery is performed in a specialist children's hospital.

Resource use and costs

Reference costs (for 2010/11) were found for hip, knee, foot, hand, shoulder and upper arm, and elbow and lower arm procedures. These procedures were classed as non-trauma, categories one and two. The reference costs were grouped by procedure and whether it was minor, intermediate or major surgery. Within these groups a weighted average cost was calculated from all procedures for categories one and two, with or without complications.

Costs ranged from £1855 (minor hand procedures) to £6241 (major hip procedures). With the children's tariff uplift these become £2969 to £9986. The average length of stay ranged from 1 day for hand procedures to 15 days for major hip procedures. Scoliosis or surgery for other spinal deformities cost on average £1797 (£2874 with uplift) and required on average 3 days in hospital.

These costs are for a finished consultant episode and so do not include rehabilitation physical therapy in the community after surgery.

Conclusion

Long-term outcomes are needed in order to develop a useful economic evaluation to assess surgery compared with no surgery. The question of timing of surgery and the need for monitoring would benefit from an economic evaluation. The increased costs of routinely monitoring children can be compared with the potential improvements in the effectiveness of surgery and reduction in need for further interventions. Further research would be useful in this area.

11.7. Selective dorsal rhizotomy

Health economic question

What is the cost effectiveness of selective dorsal rhizotomy (SDR) in children and young people with spasticity caused by a non-progressive brain disorder?

Literature review

No published health economic analyses were identified for this question. The NICE Interventional Procedures Advisory Committee (IPAC), which developed Selective dorsal rhizotomy for spasticity in cerebral palsy (NICE IPG 373, 2010), had access to an unpublished dissertation presenting a pilot economic analysis of SDR in the UK (Edwards 2010). It was developed to determine whether a full-scale economic analysis of SDR was needed and whether SDR should continue to be offered in the UK. The analysis was based on a group of patients treated at the Robert Jones and Agnes Hunt Orthopaedic and District Hospital NHS Trust who had undergone SDR and had been regular patients from age 5 years to post-adolescence and comprehensive hospital records were available. The costs and outcomes for this group were retrospectively analysed.

The comparison group was four patients with spastic diplegia who had not been selected for SDR for minimal clinical reasons. It was expected that these patients would have followed a very similar pattern of musculoskeletal development and impairment to the SDR group had they not undergone SDR.

The small number of patients included in the analysis makes the results of the economic evaluation uncertain, as was explained in the discussion of the dissertation. The literature review for this clinical question was limited. The only statistically significant benefit reported was a reduction in tone in lower extremity joints. As no good quality long-term data was available it is not possible to say from the evidence whether the initial reduction in tone reported would lead to long-term, clinically significant benefits.

Further analysis

Introduction

It was not possible to develop a cost effectiveness analysis as it is necessary to have final outcomes and the GDG was unable to extrapolate the reduction in tone to a potential long-term clinically significant improvement in function.

Methods

Evidence from an unpublished study would not normally be included in a NICE clinical guideline. However, the study provided useful estimates of the resource use and costs which the GDG was able to include in its consideration of SDR. The cost analysis developed for the dissertation (Edwards 2010) was very detailed and gives a thorough understanding of the costs involved in SDR and the number of consultations needed, and begins to look at the potential impact on need for further surgery. The costs are reproduced in this section. In order to understand how SDR fits into the NHS this cost data needs to be reviewed alongside good quality comparative long-term effectiveness data with a large enough population to capture the risks as well as the benefits.

Resource use and costs

A report for the Australian Medical Services Advisory Committee outlined the requirements for a centre to offer SDR. An experienced multidisciplinary team is necessary and a key aspect of the service is patient selection and monitoring. The surgery is performed by a paediatric neurosurgeon supported by specialists in paediatric anaesthesia, paediatric perioperative pain management, paediatric rehabilitation and intra-operative spinal monitoring. Post-operative care involves input from specialists in neurosurgery, paediatric rehabilitation, orthotics, orthopaedic surgery, physiotherapy, occupational therapy, nursing psychology and social work. The report stated that the procedure is not technically difficult and staff can be trained quickly (Medical Services Advisory Committee 2006).

A cost analysis was conducted for each patient in the Edwards 2010 study. A data collection sheet was used to record all contacts with the hospital or one of its outreach services in schools and clinics in other Trusts. Contact episodes were separately identified as outpatient appointments, multidisciplinary team sessions, gait assessments, orthotics supplies, hospital admissions, surgical or other in-patient interventions and admissions for physiotherapy top-up.

The study authors used a bottom-up approach where resource use for nine patients receiving SDR was recorded and then costs applied, rather than taking tariffs or reference costs for episodes (see Tables 11.20–11.23).

Table 11.20. Unit costs for treatment (Edwards 2010).

Table 11.20

Unit costs for treatment (Edwards 2010).

Table 11.21. Resource use and unit costs for selective dorsal rhizotomy procedure (Edwards 2010).

Table 11.21

Resource use and unit costs for selective dorsal rhizotomy procedure (Edwards 2010).

Table 11.22. Resource use and unit costs for rehabilitation on ward after selective dorsal rhizotomy (Edwards 2010).

Table 11.22

Resource use and unit costs for rehabilitation on ward after selective dorsal rhizotomy (Edwards 2010).

Table 11.23. Total costs for selective dorsal rhizotomy assessment, procedure and follow-up (Edwards 2010).

Table 11.23

Total costs for selective dorsal rhizotomy assessment, procedure and follow-up (Edwards 2010).

The study did not include the cost of additional follow-up clinic visits because all patients are followed up routinely post-surgery. The costs of ankle–foot orthosis and footwear while on the ward were not included because a high proportion of children with spastic diplegia routinely wear ankle–foot orthoses. The neurophysiological spinal monitoring equipment was treated as a sunk cost as it is used for other spinal surgery and so was not included in the costing.

The mean cost of care for the SDR patients (from age 5 years to end of adolescent growth [girls 16 years, boys 18 years]) was £67,478 (median £71,404, range £47,511 to £86,880). In the non-SDR group the mean cost of care was £63,542 (median £56,890, range £44,842 to £95,570).

The study reported all the patient contacts for each group including musculoskeletal surgery and inpatient days including top-up physiotherapy admission. They found the number of outpatient visits showed no significant variation between groups (32.9 outpatient visits for SDR group compared with 30.3 for non-SDR group). Non-SDR patients underwent an average of three periods of surgery in total and SDR patients an average of 1.9 (this includes the SDR surgery and periods of surgery after SDR), although the SDR patients spent longer in hospital (83 days compared with 57.5 in the non-SDR group). However, these are patient numbers are small.

Conclusion

The cost data presented in the dissertation was thorough and provides useful information. Again these are small patient numbers and so it would not be productive to compare the groups. There is considerable uncertainty surrounding the effectiveness of SDR. In order to provide a useful analysis for decision making we would need to understand the long-term benefits and risks of treatment compared with the next best alternative. A ’what-if‘ analysis working backwards from the NICE threshold for cost effectiveness to determine what level of effectiveness would be needed for SDR to be considered was suggested. The GDG had limited clinical experience with patients who had had SDR and did not feel able to estimate the potential efficacy of SDR in the long term compared with physiotherapy or orthopaedic surgery given the published evidence available. A ‘what-if’ analysis relies on clinical justification of the efficacy estimates to be useful for decision making and the GDG members did not feel they would be able to support any guesses they would make with confidence.

11.8. Overall conclusions

Given the lack of published evidence, further comparative research is necessary to capture benefits of interventions for spasticity in children and young people. These outcomes need to be expressed in terms of function, pain, adverse events and quality of life, ideally using the EuroQol Group's EQ-5D instrument (a child-friendly version is available) or the Health Utilities Index (which was developed for children). Although other goals are important, NICE considers evidence of cost effectiveness of interventions across all health states in a population in terms of QALYs. Without evidence of cost effectiveness expressed in terms of QALYs, it is not possible to determine whether interventions designed to benefit children and young people with spasticity represent a good use of resources when compared with other competing calls on those same resources in the NHS.

Long-term data on outcomes are needed for the ITB pump, orthopaedic surgery and SDR because these are invasive and expensive treatments with risks associated. Also, the studies should be designed to allow subgroup analysis by severity of spasticity in terms of the GMFCS and also limb involvement (hemiplegia, diplegia and quadriplegia). Studies should be designed to allow data on resource use to be collected to allow cost analysis. Cost effectiveness analysis comparing treatments for each subgroup will provide better information for making decisions.

Footnotes

‡‡‡‡

£42 per hour of client contact with a community physiotherapist, £40 with a hospital physiotherapist – the mean was used (Curtis 2011).